I have written and discussed how important it is to capture all the diagnostic metrics around your particular Key Performance Indicator (KPI). As analysts, we have a tendency to reveal too much information. We try to explain how much we know to the business and show our value. In the midst of these discussions, we forget the audience that to whom we are disseminating information. We overload our audience with knowledge, and they get overwhelmed with the information.

In the situation described above, nobody wins. The analyst is left without the satisfaction of providing the audience with data. The audience feels like they got information, but it is not what they were looking for. This is a fundamental communication problem that arises many times in the industry and we need to start from Strategy, move to Data and then create KPIs.

Strategy
Your website, landing page, microsite, and mobile page was built for a reason. There are marketing initiatives that are driving users to view the technical property. Strong analysts find out those reasons and understand the user experience of the page and it’s relationship to the data. Collecting this information is a vital step to finding your KPI.
[Read more…]

In my past life I used WebTrends 8, then some time has passed and I have moved into customizing WebTrends Analytics 10. I have seen some drastic improvements since the older version. I want to highlight four new features and how they can be beneficial to you or your organization.

Dashboards
The new WebTrends A10 has two dashboards that are provided. The system has an overall site dashboard that you can customize with important high level reports. There are also page dashboards that provide the visits, bounce rates, traffic sources, next page flow, social buzz, geography, search engines and search phrases reports. Both of these dashboards are really nice if you have an organization that looks at their site in specific sections or product categories. Your team can look at the high level pages of your products and determine where the traffic is coming from, bouncing, and where they are going. The design is easy to use and very simple to garner basic analytical insight. [Read more…]

You worked hard to write blogs that provide your audience with content they care about. Then, you took the necessary steps to promote your content via social channels (linkedin, Twitter, G+, Facebook, etc). After a few days, you start looking in your web analytics platform such as Google Analytics and understand if your content reached the right audience.

The pages report lets you know how many people viewed your blog post. Furthermore, you see visits, page exits, and bounce rates. But this doesn’t tell you if your blog is a hit with your audience. A blog pages report, is just page view by a unanimous person somewhere out there on the interwebs, and there is no way of telling if they will ever come back and read your blog or do business with your company. Keeping this in mind, here are a few creative ways to judge your blog content using Google Tools.

Location, Location, Location
Our company is located in Chicago, and most of our business is U.S. based at this time. For this reason, it is important to judge our blog content by how many users in the U.S. read the blog article. When you look at the pages report in Google Analytics, use a secondary dimension to pull up the Region for each blog post. Then, look at the Metro area and City reports to see how many people viewed the post in your local, metro area or city. [Read more…]

1. Educate their teams – analytics is complicated, make sure everyone understands the data they are looking at and where it is coming from.
2. Be approachable – it’s important to be knowledgeable, but it’s more important to be approachable. Your team needs to be able to ask you questions and get understandable answers.
3. Don’t fear the system that is collecting data – with so many software choices and options, it can be overwhelming to learn a new system. You should never fear learning a new web analytics platform, CRM System, Social Media Hub, or Mobile Platform. They all collect various data which is key to your success as an analyst.
4. Seek one version of the truth – data can be manipulated in different ways and achieve various results. Good analysts look for causal relationships in data that allows them to have high confidence in the analysis.
5. Be curious – listen to your team for key information that is going on in your organization. These can lead to interesting data findings.
6. Seek 90/10 – 90% of an analysts time should be analyzing data, 10% should be pulling data. This is a very big goal in the industry.
7. Suggest A/B Tests – sites always need improvement. A/B tests are a great way to move the needle in a positive direction. Good analysts should always be on the lookout for something to test and improve.
8. Segment – groups of people may behave a certain way with your product, mobile device, or website. Finding these segments and providing them with what they want will create revenue for your organization.
9. Document caveats – everything has a condition or something important to keep in mind while interpreting the data.
10. Work with raw data – many systems simplify the analysis for your, but that isn’t always enough to obtain statistical significance. Exporting raw data into an SPSS, SAS or R allows you to obtain insights with statistical significance.

Related Posts:
3 ways to find something to A/B Test
How to become a great web analyst
When to switch your web analytics platform

Website redesign’s happen quite frequently these days. Unfortunately, for a web analyst this isn’t a time to go on vacation. Many changes in the organizational direction, copy, site functionality happen during a website redesign. These changes greatly impact your analytics approach and tagging of your current web properties. Here are 4 tips for a web analyst to survive a website redesign.

Tip #1: Get Wireframes and Functionality Early
When you know how the new site templates are going to be laid out, it will alert you to possible changes in the way users browse the website. The changes in user path, user experience, and functionality will greatly affect your website analytics data. Depending on how sites are laid out, your engagement with time on site and pages per visit will vastly change with a redesign. The sooner you get your hands on the wireframes, the more prepared you will be to create tags that capture all the new interesting aspects of your redesigned website. [Read more…]

The web is not static, and neither are your competitors. Your web analytics team should always be looking to create hypothesis to optimize your website. Optimization is about finding areas of your site that can be improved on a systematic basis, and testing the validity of those hypothesis.

Web Analytics Hypothesis
Web analytics management needs to provide your team with the necessary software and tools to conduct proper analysis and create hypothesis. This can be mouse tracking software, deeper dive software such as Omniture Discover, survey software such as Forsee, and social measuring tools. These tools need to be used in conjunction with your web analytics platform. Providing your team with many tools to analyze pages will help them validate their hypothesis and create use cases.

Web Analytics Validation
This one is very important, as your web analytics manager should validate if the hypothesis are correct statistically and organizationally. Web managers need to validate the quality of the hypothesis by statistical means, making sure that significant data was collected. Web analytics managers also need to validate that the hypothesis aligns with organizational goals.

For example, analyst came up with the hypothesis that the Green Widget page needs to have a form on right rather than a form on left. This was done by analyzing the mouse tracking software. Web analytics manager needs to review that the eye tracking software had significant visits to create this hypothesis. More importantly, the manager needs to ask product managers if the organization is pushing to sell more Green Widgets. Any misalignment in the validation phase causes further testing, optimization and measurement for a possibly discontinued product.

Website Testing, Optimization, Measurement
Validated hypothesis need to be tested. Proper A/B or Multivariate testing techniques need to be conducted to make sure that the hypothesis will provide significant lift for your organization. Once tests are conclusive, most optimization software allows you to leave the winning version on your site until your website team has updated the code with the changes. During this process, your web analytics team should be measuring the results of the update and documenting the data for sharing.

Sharing Results With Executives, and Product Managers
Once your team has been through the optimization cycle several times, you are ready to share results. This is a very important step in driving a data driven organizational culture. Product managers and executives who see successful improvement in website performance become advocates of the structure and process. New hypothesis for tests are developed and validated. Your web analytics is now driving significant value to your organization.

Conclusion
With everything in data there are caveats. The most important caveat to the above process is your web analytics tags have to be deployed correctly and data has to be validated. Incorrect sourcing of data leads to distrust and breaking of the web optimization process.

Related Posts:
3 ways to find something to A/B Test
How to become a great web analyst
When to switch your web analytics platform

All clients, internal or external, are looking for quality insights. Definition of quality is how actionable the insights are and the positive impact the insights will produce. An analyst has to be confident in the data that is being delivered to business stakeholders. Analytical insight confidence comes from the Reliability and Validity of the insights.

Reliability
Reliability of the data comes from understanding the sources of data. It is mostly about the consistency of your data sample. Data that has a very large and diverse distribution may not be reliable for what you are looking to accomplish. Large variances in data are not consistent and will throw off the reliability of your insights. [Read more…]